Direct content-based retrieval from music scores images

TL;DR

This paper explores content-based retrieval methods for music score images, comparing transcription-based, transcription-free, and language model approaches across diverse datasets.

Contribution

It systematically evaluates multiple retrieval techniques, including a novel transcription-free Transformer model and a dataset construction method from annotated corpora.

Findings

OMR-based methods excel in in-domain retrieval

Transcription-free models better handle domain variability

Different methods perform best under different dataset conditions

Abstract

The digitization of musical scores plays a crucial role in their preservation and accessibility, yet information retrieval still depends mainly on metadata searches, such as by title or composer. Content based search in music score images remains underexplored compared to text documents, despite its potential value for musicians, musicologists, and educators. This work contributes to the field by first studying which characteristics of a score are most relevant for search and by defining a systematic method to build query datasets from any annotated corpus. We also consider diverse methods for content-based search on music score images, ranging from transcription-based approaches relying on Optical Music Recognition (OMR), to a transcription-free Transformer model trained to recognize queries directly from score images, and a text-prompted Large Language Model. Our experiments evaluate these models on four corpora exhibiting diverse characteristics in terms of dataset size, image quality, and typesetting mechanisms. Overall, each method excels under different conditions: OMR-based pipelines achieve higher in-domain retrieval, whereas transcription-free models handle domain variability more effectively.